Method and apparatus for blind source separation improving interference estimation in binaural Wiener filtering

Simple SummaryContent extracted from patent full text and abstract with AI.

This patent describes a method and system for noise reduction in binaural (two-ear) microphone setups. It uses an improved Wiener filtering technique, where interference (unwanted noise sources) is more accurately estimated through blind source separation (BSS) algorithms. The processed microphone signals keep the natural binaural cues intact, resulting in clear separation of a desired target sound from background noise, primarily benefiting users of binaural hearing aids.

Use CasesContent extracted from patent full text and abstract with AI.

• Binaural hearing aids for people with hearing impairment to better distinguish relevant speech from background noise.
• High-fidelity audio recording devices aiming to capture a specific sound source in noisy environments while preserving spatial audio cues.
• Voice communication systems, such as teleconferencing headsets, to improve clarity in multi-speaker or noisy settings.
• Assistive listening devices in public venues for enhanced speech intelligibility.
• Robotic auditory perception systems that require spatial hearing to isolate sounds in complex soundscapes.

BenefitsContent extracted from patent full text and abstract with AI.

• More precise and effective background noise and interference reduction even when there are more noise sources than microphones.
• Preservation of natural binaural (spatial) cues, improving the realism and orientation abilities for users.
• Works without needing prior information about the direction or location of the sound sources (blind source separation).
• Reduces perceptual artifacts and maintains speech clarity, thus improving user experience, especially in hearing aids.
• Enhances the performance of audio devices in both anechoic (non-reverberant) and reverberant environments.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

The invention claims a method and an appropriate acoustic signal processing system for noise reduction of a binaural microphone signal ( x 1 , x 2 ) with one target point source (s) and M interfering point sources (n 1 , n 2 , ...,n M ) as input sources to a left and a right microphone (2) of a binaural microphone system, comprising the step of: - filtering a left and a right microphone signal ( x 1 , x 2 ) by a Wiener filter (14) to obtain binaural output signals ( s L , s R ) of the target point source (s), where said Wiener filter (14) is calculated as H W = 1 - ¦ x 1 , n + x 2 , n ¢ x 1 , n + x 2 , n ¦ x 1 + x 2 ¢ x 1 + x 2 , where H w is said Wiener filter (14), ¦( x 1, n + x 2, n )( x 1,n+ x 2, n ) is the auto power spectral density of the sum of all the M interfering point sources components (x1,n, x2,n) contained in the left and right microphone signal (x1, x2) and ¦( x 1 + x 2 )( x 1 + x 2 ) is the auto power spectral density of the sum of the left and right microphone signal (x1, x2). Owing to the linear-phase property of the calculated Wiener filter (14), original binaural cues are perfectly preserved not only for the target source (s) but also for the residual interfering sources (n 1 , n 2 , ... n M ) .